摘要
燃煤电厂脱硝控制系统的调节品质对发电机组的经济性和环保性有重要的意义。本文研究了以机组负荷为时变量的脱硝系统线性参数变化(LPV)模型辨识方法。通过机组的运行历史数据进行系统辨识,得到能够描述全局动态特性的LPV模型,并采用量子粒子群优化算法进行高维参数的优化。在进化过程中引入加速度因子,提高算法的优化效率,使优化结果更逼近最优值,通过标准测试函数仿真实验验证了算法的高效性。并将该算法应用于某电厂600 MW机组的喷氨脱硝系统建模中,辨识结果表明建立的数学模型有效可行。
The regulation quality of boiler denitrification system control has great significance for unit safety and economy operation. On the basis of the time-varying parameters of the unit load, this article presents the denitrification system＇s nonlinear model identification method. The linear parameter time-varying （LPV） model was established by the historical data, which can reflect the global dynamic properties of the object. Moreover, to optimize the high-dimensional parameters, the evolutionary acceleration factor was added into the quantum particle swarm optimization algorithm. The modified algorithm＇s optimization efficiency was improved, and the optimization effect was more close to the optimal value. The high efficiency of this algorithm was verified through high-dimensional standard function test. Finally, this solution was applied to the denitrification control system modeling of a 600 MW unit, the identification results show that the denitrification conversion model is valid, and the program is feasible.
作者
袁世通
YUAN Shitong(Datang Central-China Electrie Power Test Research Institute, Zhengzhou 450000, Chin)
出处
《热力发电》
CAS
北大核心
2017年第6期94-100,共7页
Thermal Power Generation
关键词
燃煤电厂
脱硝系统
建模
网络化LPV模型
G-QPSO算法
历史数据
辨识
coal-fired power plant, denitration system, modeling, networked LPV model, G-QPSO algorithm, historical data, identification
